In the wireless communication technology, in view of the inter-code interference and the multi-user interference caused by multi-path, an advance receiver, as the downlink receiver for Wideband Code Division Multiple Access (WCDMA), utilizes the Linear Minimum Mean Square Error (LMMSE) equalization reception technology to obtain an equalization weight Wd by channel estimation and Gaussian white noise (shorthanded as white noise below) power estimation according to the principle of a minimum mean square error between an equalized signal and a transmitted signal. Then, an LMMSE equalizer eliminates the interference caused by multi-path according to the equalization weight Wd, so as to generate an optimum reception signal.
Presently, the method for estimating a Gaussian white noise power in a multi-path signal obtains the Gaussian white noise power by deriving a single-path noise through a pilot channel. In a single antenna mode, a pilot signal transmits a fixed symbol, assumed as A. In a transmit diversity mode, all the symbols transmitted by a first antenna are A, and the symbols transmitted by a second antenna are periodically repeated in a pattern of <A,-A,-A,A>, with every two pilot symbols as a group. The symbols are regrouped so that the former and latter symbols in each group are the same. In this way, the same method can be utilized for estimating the Gaussian white noise power in receiving signals transmitted by a transmit diversity and a single antenna. That is, the noise power on the signal of each path is the remaining symbols obtained by removing the signal from the subtraction result of the former and latter symbols.
During the implementation of the present disclosure, the inventors have found that the existing systems has at least the following problems: the interference exists between two paths because cross-correlation is not zero when the code is not aligned, and the noise power in the signal of each path obtained by the existing systems also contains the inter-path interference noise power instead of being a pure Gaussian white noise power. Therefore, the Gaussian white noise power needed in LMMSE equalization cannot be characterized accurately, causing performance loss of equalization.